import os
import audiomate
from audiomate import annotations
from audiomate.utils import download
from audiomate.utils import files
from . import base
DOWNLOAD_URL = 'http://opihi.cs.uvic.ca/sound/music_speech.tar.gz'
DIRECTORIES = {'music_wav': 'music', 'speech_wav': 'speech'}
[docs]class GtzanDownloader(base.CorpusDownloader):
"""
Downloader for the GTZAN Corpus.
Args:
url (str): The url to download the dataset from. If not given the default URL is used.
It is expected to be a tar.gz file.
"""
def __init__(self, url=None):
if url is None:
self.url = DOWNLOAD_URL
else:
self.url = url
[docs] @classmethod
def type(cls):
return 'gtzan'
def _download(self, target_path):
os.makedirs(target_path, exist_ok=True)
tmp_file = os.path.join(target_path, 'tmp_ark.tar.gz')
download.download_file(self.url, tmp_file)
download.extract_tar(tmp_file, target_path)
# We use copy since subfolders in the archive are read-only, hence throws permission error when trying to move.
files.move_all_files_from_subfolders_to_top(target_path, delete_subfolders=True, copy=True)
os.remove(tmp_file)
[docs]class GtzanReader(base.CorpusReader):
"""
Reader for the GTZAN music/speech corpus. The corpus consits of 64 music and 64 speech tracks that are each 30
seconds long. The Wave files are 16-bit mono and have a sampling rate of 22050 Hz.
.. seealso::
`MARSYAS: GTZAN Music/Speech <https://marsyasweb.appspot.com/download/data_sets/>`_
Download page
"""
[docs] @classmethod
def type(cls):
return 'gtzan'
def _check_for_missing_files(self, path):
return []
def _load(self, path):
corpus = audiomate.Corpus(path=path)
for directory, type_name in DIRECTORIES.items():
source_directory = os.path.join(path, directory)
if not os.path.isdir(source_directory):
continue
it = os.scandir(source_directory)
for entry in it:
if not entry.name.endswith('.wav'):
continue
file_path = os.path.join(source_directory, entry.name)
file_idx = entry.name[0:-4] # chop of .wav
utterance_idx = file_idx # every file is a separate utterance
corpus.new_file(file_path, track_idx=file_idx, copy_file=False)
utterance = corpus.new_utterance(utterance_idx, file_idx)
utterance.set_label_list(annotations.LabelList.create_single(type_name, idx=audiomate.corpus.LL_DOMAIN))
return corpus